Biochemistry 2003, 42, 2607-2615
2607
Partial Rescue of Functional Interactions of a Nonpalmitoylated Mutant of the G-Protein GRs by Fusion to the β-Adrenergic Receptor† Ozlem Ugur,*,‡ H. Ongun Onaran,‡,§ and Teresa L. Z. Jones| Department of Pharmacology and Clinical Pharmacology and Molecular Biology and Technology Research and DeVelopment Unit, Ankara UniVersity Faculty of Medicine, Sihhiye, 06100 Ankara, Turkey, and Metabolic Diseases Branch, National Institute of Diabetes and DigestiVe Diseases, National Institutes of Health, Bethesda, Maryland 20892 ReceiVed July 17, 2002; ReVised Manuscript ReceiVed December 31, 2002
ABSTRACT: Most heterotrimeric G-protein R subunits are posttranslationally modified by palmitoylation, a reversible process that is dynamically regulated. We analyzed the effects of GRs palmitoylation for its intracellular distribution and ability to couple to the β-adrenergic receptor (βAR) and stimulate adenylyl cyclase. Subcellular fractionation and immunofluorescence microscopy of stably transfected cyc- cells, which lack endogenous GRs, showed that wild-type GRs was predominantly localized at the plasma membrane, but the mutant C3A-GRs, which does not incorporate [3H]palmitate, was mostly associated with intracellular membranes. In agreement with this mislocalization, C3A-GRs showed neither isoproterenol- or GTPγS-stimulated adenylyl cyclase activation nor GTPγS-sensitive high-affinity agonist binding, all of which were present in the wild-type GRs expressing cells. Fusion of C3A-GRs with the βAR [βAR-(C3A)GRs] partially rescued its ability to induce high-affinity agonist binding and to stimulate adenylyl cyclase activity after isoproterenol or GTPγS treatment. In comparison to results with the WTGRs and βAR (βAR-GRs) fusion protein, the βAR-(C3A)GRs fusion protein was about half as efficient at coupling to the receptor and effector. Chemical depalmitoylation by hydroxylamine of membranes expressing βAR-GRs reduced the high-affinity agonist binding and adenylyl cyclase activation to a similar degree as that observed in βAR-(C3A)GRs expressing membranes. Altogether, these findings indicate that palmitoylation ensured proper localization of GRs and facilitated bimolecular interactions of GRs with the βAR and adenylyl cyclase.
Heterotrimeric G-proteins1 play a crucial role at the plasma membrane, transducing extracellular signals of various kinds to intracellular effector molecules such as adenylyl cyclase, phospholipase C, and ion channels (1, 2). Each heterotrimer consists of an R subunit, which exchanges GDP for GTP upon activation by its cognate receptor, and β and γ subunits, which are tightly bound together. In contrast to the receptors and effectors that are integral membrane proteins, G-protein subunits are peripheral membrane proteins associated with the inner face of the plasma membrane. The membrane attachment of GR and Gγ subunits is facilitated by their covalent modification with lipids (3, 4). Most GR subunits undergo palmitoylation, the posttranslational attachment of palmitic acid through a labile thioester bond to amino† This work was supported in part by grants from the Turkish Scientific and Technical Research Council (SBAG-2105) and the Ankara University Research Fund (2001-08-09-054). * To whom correspondence should be addressed. Tel: 90-312-310 6268. Fax: 90-312-310 6268. E-mail:
[email protected]. ‡ Department of Pharmacology and Clinical Pharmacology, Ankara University Faculty of Medicine. § Molecular Biology and Technology Research and Development Unit, Ankara University Faculty of Medicine. | National Institutes of Health. 1 Abbreviations: G-protein, guanine nucleotide-binding protein; βAR, β-adrenergic receptor; WT, wild type; CYP, cyanopindolol; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; BSA, bovine serum albumin; ERK, extracellular signal regulated kinase; SEM, standard error of the mean; AUC, area under curve.
terminal cysteine residues. Palmitoylation is reversible, and palmitate turnover accelerates upon G-protein activation, suggesting a regulatory role for palmitoylation in G-protein signaling (5, 6). However, the function of palmitoylation during G-protein signaling and particularly Gs signaling is not clearly understood. Does it act as a membrane-targeting signal or does it directly affect protein-protein interactions? Site-directed mutagenesis to prevent palmitoylation on GR subunits leads to defects in receptor and effector interactions (7-15), but mutation of cysteine residues, independent of the loss of palmitoylation, can alter receptor and effector interactions (13). In addition, depalmitoylation can reduce the effective concentration of GR subunits at the plasma membrane (15-17). A decrement in plasma membrane localization makes evaluation of the role of palmitoylation in receptor and effector interactions difficult because these interactions depend on both the affinity and the relative plasma membrane concentrations of these proteins. To circumvent the problems of plasma membrane targeting, receptor-GR fusion proteins have successfully been created that fuse the carboxy-terminal tail of the receptor with the amino terminus of the GR subunit (reviewed in refs 18 and 19). Fusion proteins of receptors with the nonpalmitoylated mutants of GRi1 and GR11 show a full restoration of receptor interactions for GRi1 and at least a partial restoration for GR11 (10, 20). The effect of GR palmitoylation on GR-effector interactions was not tested in these studies (10, 20). βAR-
10.1021/bi026470i CCC: $25.00 © 2003 American Chemical Society Published on Web 02/11/2003
2608 Biochemistry, Vol. 42, No. 9, 2003 GRs fusion proteins have also been created but not previously used to test the function of GRs palmitoylation. The βARGRs fusion protein can undergo palmitoylation on both the βAR and GRs (21), efficiently couple to adenylyl cyclase, maintain high- and low-affinity binding states, and show GTP turnover rates similar to rates obtained after reconstitution of purified proteins into phospholipid vesicles (18). The goal of the present study was to determine the importance of palmitoylation on GRs for its receptor and effector interactions. For this purpose, (1) we stably transfected S49 lymphoma cyc- cells, which do not express GRs, with wild-type (WT) or a nonpalmitoylated mutant of GRs alone or as fusion proteins with the βAR, and (2) we treated membranes from the cells expressing the fusion proteins with hydroxylamine to remove the palmitate. We found that palmitoylation both targets GRs to the plasma membrane and improves βAR-GRs and GRs-adenylyl cyclase interactions independent of membrane targeting. EXPERIMENTAL PROCEDURES Materials. Cell culture media, fetal bovine serum, and antibiotics were purchased from Biochrom (Berlin, Germany). Enzymes used in molecular cloning were obtained from New England Biolabs (Beverly, MA). GTP, GTPγS, ATP, Pwo polymerase, and Complete Mini protease inhibitors were purchased from Roche Diagnostics (GmbH, Manheim, Germany). Isobutylmethylxanthine, hydroxylamine hydrochloride, forskolin, and (-)-isoproterenol were purchased from Sigma (Germany). Protein A-Sepharose CL4B, [3H]dihydroalprenolol, and [125I]cyanopindolol were purchased from Amersham-Pharmacia (Buckinghamshire, England). Plasmid Constructs and Mutagenesis. The cDNA encoding the long form of rat GRs was in the pCDNA 3.1 (+) vector (Invitrogen, Carlsbad, CA). The C3A mutant of GRs was generated using a site-directed mutagenesis method based on the polymerase chain reaction, QuikChange (Stratagene, La Jolla, CA), except the Pwo polymerase (Roche Diagnostics) was used. The cDNA encoding the 822 amino acid long human β2-adrenergic receptor-rat GRs fusion protein (βARGRs) in the pCDNA3 vector was a kind gift of Dr. Tommasso Costa (Istituto Superiore de Sanita, Roma, Italy). The same mutagenesis method was used to generate a palmitoylationdeficient fusion protein where the third residue on the GRs portion, a cysteine, was replaced with alanine [βAR(C3A)GRs]. Mutations were confirmed using ABI PRISM dye terminator cycle sequencing (Perkin-Elmer-Cetus, Norwalk, CT). Cell Culture and Stable Transfection. S49 lymphoma cyccells were grown in Dulbecco’s modified Eagle’s media supplemented with penicillin (100 IU/mL), streptomycin (100 µg/mL), and 10% (v/v) fetal bovine serum, at 37 °C in a humidified atmosphere of 5% CO2. Cyc- cells were transfected by electroporation essentially as described before (22) except that transfections were performed with the Electroporator II (Invitrogen, San Diego, CA) at a setting of 500 V and 71 µF, and 40 µg of circular plasmid DNA was used for each transfection. Twenty-four hours after the transfection, cells were diluted with 20 mL of medium containing Geneticin (Gibco, Carlsbad, CA) at a final concentration of
Ugur et al. 700 µg/mL and distributed into 96-well plates in 0.2 mL aliquots. After 4 weeks, Geneticin-resistant clones were selected further according to their expression levels as assessed by immunoblotting [for the clones transfected with GRs or (C3A)GRs] or isoproterenol-induced cAMP accumulation levels [for the clones transfected with the βARGRs or βAR-(C3A)GRs]. Thus, clones expressing approximately equal amounts of GRs or (C3A)GRs in their membrane fractions and the βAR-GRs or βAR-(C3A)GRs clones having comparable isoproterenol-induced cAMP accumulation levels were selected and grown for further experiments. Immunoblotting. Proteins were separated by SDS-PAGE, transferred to nitrocellulose paper, and detected with the polyclonal, affinity-purified RM antibody, which recognizes the carboxy-terminal decapeptide of GRs (1 µg/mL) or a polyclonal antibody to the extracellular signal regulated kinase 2 (ERK2) (Santa Cruz Biotechnologies) (1:1000), and enhanced chemiluminesence (Amersham, Arlington Heights, IL) using horse radish peroxidase-conjugated goat anti-rabbit and goat anti-mouse antibodies (developed in M.B.T.R.D. Unit, Ankara University). An image analysis software (AIDA) and a camera system (Raytest, Straubenhardt, Germany) were used for the densitometric analyses of protein bands. Immunocytochemistry and Confocal Microscopy. Since cyc- cells do not attach to glass or plastic surfaces even in the presence of extracellular matrix proteins, all the incubation and washing steps were performed by resuspending the cells in the indicated solutions followed by centrifugation at 200g. Otherwise, all of the procedures were essentially as described before (23). Briefly, (1-2) × 106 cells were washed three times with 1 mL of PBS at room temperature, fixed in 2% (w/v) paraformaldehyde in PBS for 20 min, and permeabilized with 0.1% Triton X-100 (v/v) in PBS for 15 min. After being blocked with 1% (w/v) BSA in PBS, cells were incubated for 1 h with 0.5 µg/mL RM antibody in PBS containing 0.1% BSA, followed by washing three times with 0.1% BSA in PBS. Cells were then incubated for 1 h with Cy3-labeled anti-rabbit antibody (ZYMED, San Francisco, CA) diluted 1:2000 in PBS containing 0.1% BSA. After being washed three times with PBS, cells were resuspended in 50 µL of PBS, pipetted onto a microscope slide, covered, and immediately visualized with a Zeiss LSM-510 confocal microscope (Germany) equipped with 543 He-Ne laser and 40× and 63× Zeiss plan-apo oil immersion objectives. Cell Fractionation and Membrane Preparation. Cyccells, washed and pelleted in PBS, were lysed, homogenized, and fractionated into particulate and soluble fractions by centrifugation at 125000g for 1 h as described before, except mannitol was omitted from the homogenization buffer (24). For the adenylyl cyclase and binding assays, cells were lysed and homogenized by passing 10-15 times through a 26gauge syringe tip in a buffer containing 5 mM Tris-HCl, pH 7.4, and a mixture of protease inhibitors, diluted according to manufacturer’s recommendation (Complete Mini, Roche Diagnostics). Unbroken cells and nuclei were pelleted by centrifugation at 400g for 5 min. The membrane pellet, obtained by centrifugation for 1 h at 40000g of the supernatant, was washed once with a buffer containing 50 mM Tris-HCl, pH 7.4, 5 mM MgCl2, protease mixture, and 0.2 mg/mL dithiothreitol. The pellet was resuspended at a
GRs Palmitoylation and Receptor-Effector Coupling concentration of 2-3 mg/mL in a buffer containing 50 mM Tris-HCl, pH 7.4, 10 mM MgCl2, protease inhibitors, and 25% (w/v) sucrose and stored at -70 °C. Protein concentrations were determined by the Bradford assay using bovine serum albumin (BSA) as the standard (25). Membrane Attachment after GTPγS Treatment. Sixty micrograms of cell membranes was incubated with and without 10 µM GTPγS in a buffer containing 50 mM TrisHCl, pH 7.4, 100 mM NaCl, and 10 mM MgCl2 for 4 h at 37 °C. The samples were separated into particulate and soluble fractions by centrifugation at 125000g for 1 h. The pellets were resuspended in the initial volume of buffer. Equal volumes of particulate and soluble fractions were subjected to SDS-PAGE for immunoblotting. Determination of Membrane Adenylyl Cyclase ActiVity and cAMP Accumulation in Cells. Adenylyl cyclase activity in cell membranes was measured essentially as described before (26). Briefly, membranes were preincubated in 96-well plates (0.5-1 µg/well) with indicated stimulants for 5 min (or at least 1.5 h with GTPγS) in 100 µL, at 37 °C. Assay was initiated by adding 50 µL of adenylyl cyclase assay buffer (final concentrations: 50 mM Tris-HCl, pH 7.4, 100 mM KCl, 10 mM MgCl2, 0.5 mM ATP, and 1 mM isobutylmethylxanthine) and terminated after 5 min by adding 50 µL of 0.4 N HCl. For cAMP accumulation experiments in whole cells, cyc- cells were washed with serum-free media, incubated with 1 mM isobutylmethylxanthine for 15 min and with isoproterenol for an additional 15 min. Reaction was terminated by addition of 0.5 volume of 0.4 N HCl. The cAMP amount was measured by radioimmunoassay as described before (26). Binding Experiments. Three micrograms of the membranes expressing WT- or C3A-GRs or 1.5 µg of the membranes expressing βAR-GRs or βAR-(C3A)GRs was incubated in the presence of 20-40 pM [125I]CYP and the indicated concentrations of isoproterenol, with or without GTPγS (10 µM) in the binding buffer (50 mM Tris-HCl, pH 7.4, 100 mM KCl, 10 mM MgCl2) in a final volume of 250 µL, at 27 °C for 3 h. Pipettings were performed using a Biomek 2000 laboratory automation workstation (Beckman Instruments, Fullerton, CA). Reaction was stopped by filtration through Whatman GF/B filters with the use of a cell harvester (Skatron, Lier, Norway). Data were analyzed using the LIGAND program (27) or by the regression of the fourparameter logistic equation. GTPγS-induced shift in isoproterenol binding curves was analyzed by evaluating the area between the two binding curves obtained in the presence or absence of GTPγS. The area between the curves was obtained by subtracting the areas under the normalized binding curves, which were calculated according to Simpson’s rule of numerical integration. Membrane amounts of the fusion proteins were determined by single point saturation binding assays, using 2 µg of membrane protein and 2 nM [125I]CYP. Nonspecific binding was estimated in the presence of 10 µM propranolol. Whole cell saturation binding experıments were carried out by using [3H]dihydroalprenolol as the radioligand at a concentration range of 0.003-100 nM (97 Ci/mmol specific activity). Cells that were washed with serum-free DMEM twice were incubated at 37 °C for 1 h, distributed to a 96-well plate at a concentration of 100000 cells per well, and incubated with indicated concentrations
Biochemistry, Vol. 42, No. 9, 2003 2609
FIGURE 1: Subcellular fractionation of WT- and C3A-GRs. Cyccells, stably transfected with the WT- and C3A-GRs cDNAs were separated into particulate (P) and soluble (S) fractions by centrifugation at 125000g, and 5 µg of protein from each fraction was subjected to SDS-PAGE. Proteins were analyzed by immunoblotting with the RM antibody specific for GRs or an antibody specific for the cytosolic protein, ERK2.
of [3H]dihydroalprenolol in a final volume of 100 µL. Alprenolol at a concentration of 1 µM was used to determine the nonspecific binding for each concentration of [3H]dihydroalprenolol. After incubation at 37 °C for 120 min, cells were filtered through a Whatman GF/C filter, using a cell harvester. Radioactivity on the filter was counted by liquid scintillation by the use of a Wallac Microbeta Trilux counter. Hydroxylamine Treatment. The pH of the hydroxylamine solution was adjusted to 7.4 with 100 mM Tris-HCl and KOH. The control solution contained 100 mM Tris-HCl and equimolar KCl. We observed that agonist binding characteristics of the βAR-GRs fusion protein were sensitive both to the high salt concentration present in the matching control solution of 1 M hydroxylamine (usual concentration used to depalmitoylate proteins) and to the washing step that followed the incubation. Thus, we used 0.5 M hydroxylamine and omitted the washing step. Membranes were incubated in the solutions with or without hydroxylamine for 45 min at 20 °C, immediately diluted after the incubation, and added to the binding assay. The final hydroxylamine concentration in the assay was less than 2 mM. Hydroxylamine (1 M) or the corresponding control solution (see above) was used to treat membranes for the adenylyl cyclase assays. RESULTS Membrane Attachment and Intracellular Distribution of WT- and C3A-GRs. To assess the function and intracellular location of a nonpalmitoylated mutant of GRs, we expressed the WT and C3A mutant of GRs by stable transfection of cyc- cells, which do not have endogenous GRs (28). Mutation of the cysteine at residue 3 abolishes palmitoylation of GRs (7, 24, 29). The cells were separated into crude membrane and cytosol fractions by centrifugation. ERK2, a cytosolic protein (30), was found predominantly in the soluble fraction with this fractionation procedure (Figure 1). Both the WTand the C3A-GRs were expressed primarily and approximately equally in the particulate fractions of the transfected cyc- cells with a 3-fold higher expression of C3A-GRs in the soluble fraction (Figure 1). Immunofluorescence confocal microscopy on the stably transfected cyc- cells showed that WT-GRs was primarily localized at the plasma membrane, whereas C3A-GRs was expressed within the cell (Figure 2). Essentially no staining was seen for the vector-transfected cells with the antibody specific for GRs or for the transfected cells without the specific antibody (data not shown). This result, in conjunction with the fractionation studies showing C3A-GRs predominantly in the particulate fraction (Figure 1), indicates that
2610 Biochemistry, Vol. 42, No. 9, 2003
Ugur et al.
Table 1: Functional Assays of Cyc- Cells Expressing the WT- and C3A-GRsa cAMP accumulation [pmol (106 cells)-1 min-1]b vector C3A-GRs WT-GRs
isoproterenol binding Kd (nM)d
adenylyl cyclase activity [pmol (mg of protein)-1 min-1]c
GTPγS (-)
isoproterenol (100 µM)
isoproterenol (100 µM)
GTPγS (100 µM)
KH
KL
GTPγS (+)
0.001 ( 0.006 0.42 ( 0.04 2.6 ( 0.02
5.1 ( 0.3 4.5 ( 1.2 59.4 ( 16.1
11.8 ( 1.7 20.3 ( 0.1 135.7 ( 32
4.4 ( 0.3e
116 ( 62.9 566 ( 163
163 ( 64.7 156 ( 95.4
a Cyc- cells were stably transfected with vector alone or with the cDNAs for the WT or C3A mutant of GRs. b cAMP accumulation assays were performed in whole cells after preincubation in 1 mM isobutylmethylxanthine for 15 min and then with and without isoproterenol for 15 min. Shown are the net isoproterenol-induced accumulations, i.e., the difference between the values with and without isoproterenol treatment. Data are the mean ( SEM from two experiments performed in six replicates. c The adenylyl cyclase assay was performed on membranes preincubated for 5 min with isoproterenol and 1 µM GTP or for 90 min with GTPγS and incubated in the assay buffer for 5 min as described under Experimental Procedures. Shown are the net activities, i.e., the difference between the values with and without the indicated treatments. Data are the mean ( SEM from two experiments. d The competition binding assays were performed as described under Experimental Procedures on membranes incubated with 20 pM [125I]CYP and a range of isoproterenol concentrations in the presence or absence of 10 µM GTPγS for 3 h. Binding parameters were estimated by the use of the LIGAND program. KH and KL are the high- and low-affinity equilibrium dissociation constants for isoproterenol, respectively. Data are the mean ( SEM for two experiments. e The percentage of high-affinity binding sites was 54 ( 16%.
FIGURE 3: GTPγS-induced release of GRs from the membrane. Equal amounts of membranes expressing WT- or C3A-GRs were incubated with or without 10 µM GTPγS at 37 °C for 4 h and separated into particulate and soluble fractions by centrifugation at 100000g for 1 h. Pellets were resuspended in the initial volume of buffer, and equal volumes from each fraction were subjected to immunoblotting with an antibody to GRs.
FIGURE 2: Intracellular localization of WT- and C3A-GRs in stably transfected cyc- cells. Cells expressing WT-GRs (a) and C3A-GRs (b, c) were fixed in 2% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained by incubations first with the rabbit polyclonal RM antibody to GRs and subsequently with the CY3labeled anti-rabbit antibody. A confocal microscope was used to examine the cells.
mutation of C3 led to a loss of the specific targeting of GRs to the plasma membrane but that the C3A mutant can still nonspecifically attach to intracellular membranes. C3A-GRs Cannot Functionally Interact with the βAR or Adenylyl Cyclase. We tested the ability of the nonpalmitoylated C3A-GRs to couple to the βAR and adenylyl cyclase. In a whole cell assay, the C3A-GRs transfected cells showed a minimal increase in cAMP accumulation in response to isoproterenol compared to the response in the WT-GRs transfected cells (Table 1). The adenylyl cyclase activity in response to isoproterenol or GTPγS of membranes from cells expressing the nonpalmitoylated C3A mutant was concentration-independent (data not shown) and did not differ (isoproterenol) or differed only marginally (GTPγS) from results from membranes from vector-transfected cells (Table 1). In comparison, the cells expressing the WT-GRs showed a prominent and concentration-dependent increase in adenylyl cyclase activity after incubation with isoproterenol or GTPγS
(Table 1; only the net isoproterenol- or GTPγS-induced activities are shown). We studied the agonist binding characteristics of the βAR in membranes from WT- or C3A-GRs transfected cells because the presence of high-affinity agonist binding and its sensitivity to a guanine nucleotide reflect the functional interaction between a receptor and a G-protein. The membranes from cells expressing the WT-GRs showed high- and low-affinity agonist binding, and GTPγS abolished the highaffinity component of the binding (Table 1). In membranes expressing C3A-GRs, isoproterenol exhibited single-site binding in both the absence and presence of GTPγS (Table 1). The C3A-GRs mutant was able to bind GTPγS because incubation of membranes from the stably transfected cells with GTPγS showed a release of the C3A-GRs mutant to the soluble fraction that was considerably higher compared to the release of WT-GRs or C3A-GRs without GTPγS (Figure 3). These results show that the nonpalmitoylated C3A-GRs did not functionally interact with its receptor or effector. Expression of the βAR-GRs Fusion Proteins. We assessed the role of GRs palmitoylation for βAR-Gs and Gs-adenylyl cyclase interactions independently from its role in membrane attachment and trafficking by using a βAR-GRs fusion protein (Figure 4A). This fusion protein ensures a 1:1 stoichiometry for βAR and GRs at the membrane, regardless of the trafficking properties of the GRs. The βAR-GRs fusion protein can undergo palmitoylation through a thioester linkage to residues corresponding to C341 in the βAR and C3 in GRs (21). (We were unsuccessful in immunoprecipitating adequate quantities of the fusion protein in cyc- cells to detect [3H]palmitate incorporation.) Both the βAR-GRs and the βAR-(C3A)GRs cells were expressed as approxi-
GRs Palmitoylation and Receptor-Effector Coupling
FIGURE 4: (A) Schematic representation of the fusion protein of human β2-adrenergic receptor (hβ2AR) and the long form of rat GRs (rGRs). (B) Expression of βAR-GRs and βAR-(C3A)GRs in stably transfected cyc- cell membranes. Ten micrograms of membrane protein from cyc- cells or cells stably transfected with the indicated cDNAs was subjected to SDS-PAGE and immunoblotting with an antibody specific for the carboxy terminus of GRs. The migration of the 98 kDa molecular mass marker is shown to the left. (C) Cell surface expression levels of fusion proteins. Cyccells expressing βAR-GRs (b) or βAR-(C3A)GRs (O) were incubated with the indicated concentrations of [3H]dihydroalprenolol for 2 h at 37 °C and filtered through a Whatman GF/C filter. The first four data points at lower concentrations are presented in a Scatchard graph (inset). Estimated expression levels were 10000 and 20000 copies per cell for βAR-GRs and βAR-(C3A)GRs, respectively. (D) Intracellular localization of βAR-GRs. Cells expressing βAR-GRs were fixed in 2% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained by incubations first with the rabbit polyclonal RM antibody and subsequently with the CY3-labeled anti-rabbit antibody. A confocal microscope was used to examine the cells.
mately 105 kDa proteins in stably transfected cyc- cells as seen by immunoblotting with an antibody to the carboxyterminal sequence of GRs (Figure 4B). Their expression levels as assessed by [125I]CYP saturation binding were 4.9 and 8.0 pmol/mg of membrane protein for the βAR-GRs and βAR-(C3A)GRs cells, respectively, that correlates with the higher expression level of the βAR-(C3A)GRs seen on
Biochemistry, Vol. 42, No. 9, 2003 2611 immunoblotting. Scatchard analysis of the saturation [3H]dihydrolprenolol binding experiments conducted on intact cells showed about a 2-fold higher cell surface expression of the βAR-(C3A)GRs compared to the βAR-GRs (Figure 4C). The different levels of expression were likely due to selecting clones with similar cAMP accumulation after isoproterenol stimulation (see below). The cell surface expression of the βAR-GRs was also seen by immunofluorescence microscopy using an antibody to the carboxy terminus of GRs (Figure 4D). The pattern of staining at the cell periphery for the receptor-GRs fusion protein was similar to the staining for the WT-GRs (Figure 2a). Receptor and Effector Coupling of the βAR-(C3A)GRs Fusion Protein. Cyc- cell membranes expressing the βARGRs exhibited GTPγS-sensitive, high-affinity isoproterenol binding, indicating a functional coupling of these proteins (Figure 5A). This high-affinity isoproterenol binding was present in the membranes expressing the βAR-(C3A)GRs as well but was reduced compared to the membranes expressing βAR-GRs. As a rough measure of GTPγS-induced shift in agonist binding curves, we used the area between the binding curves obtained in the presence or absence of GTPγS. This quantity depends on the relative positions of the two binding curves and was calculated as the difference between the whole areas under the two individual normalized binding curves (denoted as ∆AUC). The ∆AUC calculated from the isoproterenol binding curves of βAR-(C3A)GRs expressing membranes was about one-third of that obtained with the membranes expressing the βAR-GRs (Figure 5B). Consistent with the binding data, isoproterenol-stimulated adenylyl cyclase activity of the membranes expressing the βAR-(C3A)GRs was reduced about 45% compared to the membranes expressing βAR-GRs (Figures 6A and 7A). Maximum GTPγS-induced adenylyl cyclase activity of βAR(C3A)GRs was also reduced by about 40% compared to βAR-GRs (Figures 6B and 7B). The higher basal activity in the isoproterenol assay compared to the GTPγS assay was likely due to the presence of GTP in the assay buffer with isoproterenol. These results demonstrate that the functional interactions of C3A-GRs with the βAR or adenylyl cyclase were less efficient compared to WT-GRs when both were expressed as fusion proteins with the βAR. Hydroxylamine Treatment of Membranes from Cells Expressing the βAR-GRs Fusion Protein. To determine if our results with the βAR-(C3A)GRs fusion protein were due to mutation of the cysteine residue rather than the loss of palmitoylation, we performed functional studies on cell membranes that had been treated with hydroxylamine to break thioester bonds and release palmitate. Both partners of the βAR-GRs fusion protein can be depalmitoylated upon treatment with hydroxylamine (21). So, a possible confounding factor in our experiments could be the removal of palmitate on the βAR. However, depalmitoylation of the βAR by hydroxylamine treatment and then reconstitution into a membrane preparation did not change adenylyl cyclase activity after isoproterenol stimulation, indicating that the ligand binding and G-protein coupling characteristics of the nonpalmitoylated βAR remain intact (31). Hydroxylamine treatment of membranes from cells expressing βAR-GRs reduced the GTPγS-induced shift in isoproterenol binding by about 50% compared to membranes incubated in the control solution (Figure 5C). Treatment of
2612 Biochemistry, Vol. 42, No. 9, 2003
Ugur et al.
FIGURE 6: Isoproterenol- or GTPγS-stimulated adenylyl cyclase activity of cyc- cell membranes expressing the fusion proteins. (A) Membranes expressing βAR-GRs (b) or βAR-(C3A)GRs (O) were preincubated for 5 min with the indicated concentrations of isoproterenol. Adenylyl cyclase assay was initiated by the addition of assay buffer containing 1 µM GTP. (B) Membranes were preincubated with the indicated concentrations of GTPγS for 90 min, and then the adenylyl cyclase assay was performed in the presence of GTPγS. Shown are the representatives of two experiments performed in quadruplicate. A four-parameter logistic equation was fit to the data.
FIGURE 5: GTPγS-induced shift in isoproterenol binding in cyccell membranes. (A, B) An isoproterenol-[125I]CYP competition binding assay was performed as described in Table 1 for cyc- cell membranes expressing βAR-GRs and βAR-(C3A)GRs or (C) in membranes expressing βAR-GRs, treated with or without 0.5 M neutral hydroxylamine (Hyd). (B, C) The area between the normalized binding curves obtained in the presence or absence of GTPγS (∆AUC) was calculated as described under Experimental Procedures. Mean ∆AUCs ((SEM) of five to six experiments are presented relative to βAR-GRs in (B) and relative to Hyd (-) in (C). The asterisk notes significant differences with p < 0.05 on a paired t-test.
membranes with hydroxylamine decreased the isoproterenoland GTPγS-stimulated adenylyl cyclase activity of the βAR-
GRs fusion protein by 60% and 55%, respectively, to levels similar to corresponding activities of both the hydroxylaminetreated and untreated βAR-(C3A)GRs expressing membranes (Figure 7). Forskolin-stimulated adenylyl cyclase activities of the membranes expressing βAR-GRs and βAR-(C3A)GRs were reduced by about 25% and 35%, respectively, after hydroxylamine treatment (Figure 8), possibly from a direct effect of hydroxylamine on adenylyl cyclase (32). However, the excess 35% and 30% decrease seen after hydroxylamine treatment in isoproterenol- and GTPγS-stimulated adenylyl cyclase activities, respectively, of the βAR-GRs expressing membranes was absent in βAR-(C3A)GRs membranes (Figure 8). Hydroxylamine can have nonspecific effects including changes in membrane viscosity, but the lack of further functional impairment of the βAR-(C3A)GRs mutant beyond the nonspecific effects on adenylyl cyclase suggests that the primary effect of hydroxylamine in these experiments was depalmitoylation of GRs. These results, in conjunction with
GRs Palmitoylation and Receptor-Effector Coupling
Biochemistry, Vol. 42, No. 9, 2003 2613
FIGURE 8: Effect of hydroxylamine treatment on the adenylyl cyclase activity of cyc- cell membranes expressing the fusion proteins. Membranes were preincubated with or without 1 M hydroxylamine at room temperature for 45 min and assayed for forskolin-stimulated (black bars), isoproterenol-stimulated (white bars), or GTPγS-stimulated (gray bars) adenylyl cyclase activity as described for Figure 6. Forskolin, isoproterenol, and GTPγS concentrations were 0.1 mM, 100 µM, and 1 µM, respectively. Adenylyl cyclase activities of hydroxylamine-treated membranes are presented relative to the control membranes. Data are the mean ( SEM of four experiments performed in quadruplicate.
FIGURE 7: Effect of hydroxylamine treatment on isoproterenol- or GTPγS-stimulated adenylyl cyclase activity of cyc- cell membranes expressing the fusion proteins. Membranes were treated with (white bars) or without (black bars) 1 M hydroxylamine at room temperature for 45 min and assayed for isoproterenol-induced (A) or GTPγS-induced (B) adenylyl cyclase activity. The adenylyl cyclase assay was performed exactly as described for Figure 6. For (A), the adenylyl cyclase activity shown is the difference between the activity in the presence of 100 µM isoproterenol + 1 µM GTP and the activity in the presence of 1 µM GTP. For (B) GTPγS was used at a concentration of 1 µM. Data are the mean ( SEM of four experiments performed in quadruplicate.
those using the βAR-(C3A)GRs fusion protein, indicate that localizing a nonpalmitoylated GRs at the plasma membrane could restore some of its signaling ability. However, palmitoylation was still necessary for GRs to couple to the βAR and adenylyl cyclase with full efficiency. DISCUSSION Cells control protein palmitoylation to regulate the membrane attachment, intracellular targeting, and conformation of proteins. Here we dissected the role of palmitoylation on protein interactions for GRs. We found that GRs needed palmitoylation for plasma membrane localization. Forcing GRs to the plasma membrane through fusion of a nonpalmitoylated mutant of GRs to the βAR only partially rescued its ability to functionally couple to the βAR and adenylyl cyclase. This result is unlikely to be secondary to the cysteine point mutation because removal of palmitate from WT-GRs
showed similar results: compromised formation of a receptor-G-protein complex and poor stimulation of adenylyl cyclase. Palmitoylation Supports Plasma Membrane Localization of GRs. G-protein signaling occurs at the plasma membrane where G-proteins transduce signals from cell surface receptors to membrane-bound effectors. Palmitoylation is a critical factor in this localization of G-proteins. Mutation of aminoterminal cysteine residues that prevent palmitoylation disrupts the plasma membrane localization of some GR subunits (15-17). The intracellular location of the C3A mutant of GRs could not be determined in a previous study because it was predominantly soluble (7, 17). Our results show that the plasma membrane targeting of GRs was disrupted with mutation of the acylation site. The inability of the nonpalmitoylated C3A-GRs to talk to its receptor and effector was likely due to both decreased plasma membrane localization and the diminished protein interactions of this mutant. Protein acyl transferase activity (33) and Gβγ subunits are found at the plasma membrane (34), where they regulate reciprocally the localization of GRs: Gβγ binding is needed for palmitoylation (17), and palmitoylation increases the affinity of the GRs subunit for Gβγ (35). In addition, the hydrophobicity of palmitoylation efficiently stops translocation off the membrane (36). For the “dock and lock” model of GR targeting, Gβγ binding docks the GR subunit, and palmitoylation locks it in place (37). Therefore, our finding that the C3A mutant was located on membranes besides the plasma membrane suggests that a protein interaction besides Gβγ binding or another modification may be involved. Strong experimental evidence points to the presence of a second hydrophobic modification near the amino terminus of GRs, though it has not been identified, yet (38). This factor alone may allow weak nonspecific membrane attachment whose detection would be sensitive to experimental conditions such as epitope tagging and methods of depalmito-
2614 Biochemistry, Vol. 42, No. 9, 2003 ylation (mutational vs chemical and enzymatic) and explain the wide range of reported loss of GRs membrane attachment induced by depalmitoylation (7, 24, 29, 39). Palmitoylation of GRs and Receptor and Effector Interactions. Palmitoylation is not an absolute requirement for Gs signaling because recombinant GRs purified from bacteria is not palmitoylated and can couple to the βAR and adenylyl cyclase after reconstitution into membranes (40). Our results show that placing a nonpalmitoylated GRs close enough to the βAR and adenylyl cyclase by fusing GRs to the carboxyterminal tail of the βAR was adequate to restore some of Gs signaling. However, even with GRs residing on the correct membrane, the lack of palmitoylation still limited receptor and effector coupling. Impairment in functional GRs-receptor and GRs-effector coupling could not occur through release of GRs from the membrane because it was still fused to the receptor but could occur through changes on GRs relating to its (1) conformation, (2) proximity to the membrane, or (3) orientation to the membrane and membrane-bound proteins. These mechanisms are not mutually exclusive. Palmitoylation can also target proteins, including G-proteins, to membrane microdomains enriched in cholesterol and sphingolipids (41). However, the functional significance of this targeting for Gs signaling (42) and whether the βAR-GRs fusion protein is in these microdomains is not clear. The juxtaposition of GRs at the membrane in a limited range of orientations may be necessary for GRs to nestle in the intracellular loops of the receptor and effector. The addition of a myristoylation signal to the amino terminus of a nonpalmitoylated mutant of GRs can restore its effector coupling (7). The ability of the nonpalmitoylated GRi1 to efficiently couple to the receptor after fusion to the R2A-adrenoreceptor (20) may be explained by a closer proximity to the membrane. The carboxy-terminal tail of the R2A-adrenoreceptor has only 24 residues, whereas the βAR has 86 residues and an additional 15 residues in our construct. The crystal structure of GRs binding to the catalytic domains of adenylyl cyclase (43) and a model of G-protein coupling to the receptor, based on the crystal structure of rhodopsin and transducin (44), highlight the importance of proper orientation and distance of GRs to the membrane for these interactions. In the present study, we did not address the interaction between GRs and the Gβγ complex. Palmitoylation on GRs can increase its affinity for Gβγ (35), and a decrease in heterotrimer formation could decrease the receptor coupling of Gs. However, Gβγ complexes may be unnecessary for high-affinity agonist binding to the receptor because overexpression of Gβγ subunits in Sf9 cells did not increase the ternary complex formation of the βAR-GRs fusion protein (45). Clearly, this point needs further clarification. We also did not address palmitate turnover on the βAR-GRs fusion protein. Within a membrane preparation, palmitate turnover may not occur because an identified acyl protein thioesterase, APT1, is a cytosolic protein (46). Our results could overemphasize the importance of palmitoylation, if persistent palmitoylation of GRs improves its activity. If depalmitoylation does occur during activation or preparation of the membranes, then our results would underestimate the effect of palmitoylation because the stoichiometry of palmitoylation on GRs, which is quite high in the cell (47), could be lower
Ugur et al. depending on the enzyme activities of the thioesterase and transferase in the membrane. Agonist binding to a Gs-coupled receptor starts both a GTPase and thioacylation cycle whose actions finely regulate GRs signaling within the cell. The novel finding of this study is that palmitoylation may have a role in the functional interactions of GRs with both the βAR and adenylyl cyclase, in addition to localizing GRs at the plasma membrane. The question of whether palmitoylation of GRs modifies βAR and adenylyl cyclase interactions through a change in protein conformation or stable anchorage in a proper orientation remains to be answered. ACKNOWLEDGMENT We thank Dr. Alp Can (Histology and Embryology Department, Ankara University Faculty of Medicine) for assistance with the confocal microscopy, Dr. Yukiko Miura for help with the cyc- cells, and Dr. Graeme Milligan for critical reading of the manuscript. REFERENCES 1. Gilman, A. G. (1987) Annu. ReV. Biochem. 56, 615-649. 2. Hamm, H. E. (1998) J. Biol. Chem. 273, 669-672. 3. Casey, P. J. (1995) Science 268, 221-225. 4. Wedegaertner, P. B. (1998) Biol. Signals Recept. 7, 125-135. 5. Mumby, S. M. (1997) Curr. Opin. Cell Biol. 9, 148-154. 6. Dunphy, J. T., and Linder, M. E. (1998) Biochim. Biophys. Acta 1436, 245-261. 7. Wedegaertner, P. B., Chu, D. H., Wilson, P. T., Levis, M. J., and Bourne, H. R. (1993) J. Biol. Chem. 268, 25001-25008. 8. Edgerton, M. D., Chabert, C., Chollet, A., and Arkinstall, S. (1994) FEBS Lett. 354, 195-199. 9. Wise, A., Grassie, M. A., Parenti, M., Lee, M., Rees, S., and Milligan, G. (1997) Biochemistry 36, 10620-10629. 10. Stevens, P. A., Pediani, J., Carrillo, J. J., and Milligan, G. (2001) J. Biol. Chem. 276, 35883-35890. 11. Ponimaskin, E., Harteneck, C., Schultz, G., and Schmidt, M. F. (1998) FEBS Lett. 429, 370-374. 12. Ponimaskin, E., Behn, H., Adarichev, V., Voyno-Yasenetskaya, T. A., Offermanns, S., and Schmidt, M. F. (2000) FEBS Lett. 478, 173-177. 13. Hepler, J. R., Biddlecome, G. H., Kleuss, C., Camp, L. A., Hofmann, S. L., Ross, E. M., and Gilman, A. G. (1996) J. Biol. Chem. 271, 496-504. 14. Jones, T. L. Z., and Gutkind, J. S. (1998) Biochemistry 37, 31963202. 15. Bhattacharyya, R., and Wedegaertner, P. B. (2000) J. Biol. Chem. 275, 14992-14999. 16. Morales, J., Fishburn, C. S., Wilson, P. T., and Bourne, H. R. (1998) Mol. Biol. Cell 9, 1-14. 17. Evanko, D. S., Thiyagarajan, M. M., Siderovski, D. P., and Wedegaertner, P. B. (2001) J. Biol. Chem. 276, 23945-23953. 18. Seifert, R., Wenzel-Seifert, K., and Kobilka, B. K. (1999) Trends Pharmacol. Sci. 20, 383-389. 19. Milligan, G. (2000) Trends Pharmacol. Sci. 21, 24-28. 20. Wise, A., and Milligan, G. (1997) J. Biol. Chem. 272, 2467324678. 21. Loisel, T. P., Ansanay, H., Adam, L., Marullo, S., Seifert, R., Lagace, M., and Bouvier, M. (1999) J. Biol. Chem. 274, 3101431019. 22. Gonzales, J. M., O’Donnell, J. K., Stadel, J. M., Sweet, R. W., and Molinoff, P. B. (1992) J. Neurochem. 58, 1093-1103. 23. Ugur, O., and Jones, T. L. Z. (2000) Mol. Biol. Cell 11, 14211432. 24. Degtyarev, M. Y., Spiegel, A. M., and Jones, T. L. Z. (1993) Biochemistry 32, 8057-8061. 25. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 26. Ugur, O., and Onaran, H. O. (1997) Biochem. J. 323, 765-776. 27. Munson, P. J., and Rodbard, D. (1980) Anal. Biochem. 107, 220239. 28. Bourne, H. R., Beiderman, B., Steinberg, F., and Brothers, V. M. (1982) Mol. Pharmacol. 22, 204-210.
GRs Palmitoylation and Receptor-Effector Coupling 29. Mumby, S. M., Kleuss, C., and Gilman, A. G. (1994) Proc. Natl. Acad. Sci. U.S.A. 91, 2800-2804. 30. Hochholdinger, F., Baier, G., Nogalo, A., Bauer, B., Grunicke, H. H., and Uberall, F. (1999) Mol. Cell. Biol. 19, 8052-8065. 31. Moffett, S., Adam, L., Bonin, H., Loisel, T. P., Bouvier, M., and Mouillac, B. (1996) J. Biol. Chem. 271, 21490-21497. 32. Mollner, S., Beck, K., and Pfeuffer, T. (1995) FEBS Lett. 371, 241-244. 33. Dunphy, J. T., Greentree, W. K., Manahan, C. L., and Linder, M. E. (1996) J. Biol. Chem. 271, 7154-7159. 34. Denker, S. P., McCaffery, J. M., Palade, G. E., Insel, P. A., and Farquhar, M. G. (1996) J. Cell Biol. 133, 1027-1040. 35. Iiri, T., Backlund, P. S., Jr., Jones, T. L. Z., Wedegaertner, P. B., and Bourne, H. R. (1996) Proc. Natl. Acad. Sci. U.S.A. 93, 1459214597. 36. Yang, S., Zhang, L., and Huang, Y. (2001) FEBS Lett. 498, 7681. 37. Fishburn, C. S., Pollitt, S. K., and Bourne, H. R. (2000) Proc. Natl. Acad. Sci. U.S.A. 97, 1085-1090. 38. Kleuss, C., and Gilman, A. G. (1997) Proc. Natl. Acad. Sci. U.S.A. 94, 6116-6120.
Biochemistry, Vol. 42, No. 9, 2003 2615 39. Huang, C., Duncan, J. A., Gilman, A. G., and Mumby, S. M. (1999) Proc. Natl. Acad. Sci. U.S.A. 96, 412-417. 40. Graziano, M. P., Freissmuth, M., and Gilman, A. G. (1989) J. Biol. Chem. 264, 409-418. 41. Resh, M. D. (1999) Biochim. Biophys. Acta 1451, 1-16. 42. Miura, Y., Hanada, K., and Jones, T. L. Z. (2001) Biochemistry 40, 15418-15423. 43. Tesmer, J. J., Sunahara, R. K., Gilman, A. G., and Sprang, S. R. (1997) Science 278, 1907-1916. 44. Hamm, H. E. (2001) Proc. Natl. Acad. Sci. U.S.A. 98, 48194821. 45. Seifert, R., Lee, T. W., Lam, V. T., and Kobilka, B. K. (1998) Eur. J. Biochem. 255, 369-382. 46. Duncan, J. A., and Gilman, A. G. (1998) J. Biol. Chem. 273, 15830-15837. 47. Jones, T. L. Z., Degtyarev, M. Y., and Backlund, P. S., Jr. (1997) Biochemistry 36, 7185-7191. BI026470I